BiGnOsE

Had a couples DTC codes on my 2012 USDM Ralliart.
I have read the codes with Torque (Android) and called my dealer about it.

They warned me to never clear the DTC using anything else than their tool (MUT-III?).

The basicaaly told me that clearing DTC also clear a timing chain elongation value stored in the memory. They said that their tool backup up the value before clearing and puts it back or something like that.

I called BS on that and checked with another dealer, the told me the exact same thing. Asked the directly to the tech that worked on my car, told me the same thing.

Questions are :
1 - Any proof of that ?
2 - Anyway to do it by ourself ?
3 - If true... that means we are bound to for any clearing of DTC post warranty ?
4 - What about thoses who tune their car ?

I'd like to know what you guys think or know about it.
I'm pretty sure any parameter on the timing chain is something important not to loose it's value.

Thanks for giving your input.

BiGnOsE

richardjh

That IS interesting.

I'll see if I can find something in the ECU code...

Rich

Veronica

iTrader: (2)

Thanks Rich!

BiGnOsE

Yeah, Thanks !

Meanwhile, I found a little something on here http://www.evoxforums.com/forums/showthread.php?t=28953

Its a little off topic, but points in the same direction.

Any toughts ?

deprydation

iTrader: (1)

After reading that entire thread, the only question that was never answered to me was if reflashing the ECU will clear the learned value...

If it does, damn...that blows...

BiGnOsE

The last post on the thread is from early 2010 and, in my opinion, does not give any clear answer about what really reset the timing chain elongation value. They suggest that the clear DTC might only reset it on some specific codes.

What do you understand from that post rich?

I might be missing some little details since English is not my primary language.

Thanks

Veronica

iTrader: (2)

From the post you cite - i gather no - it does not reset unless you try to for that DTC...

But - i defer to Rich who is The Man.

richardjh

Well, I've made some progress. No actual testing, yet.


I reckon I've found the function that sets (or clears) "timing chain stretch" P0012. It's pretty compact and straightforward. It does a bit of subtractive maths on data grabbed from RAM addresses, and compares the results against some constant values.

There seem to be two key values. If both go above the trigger value, P0012 is set.


The next step is for me to log these special RAM addresses, and record their values...

Then clear DTCs, and see if they are blatted.

Then reflash, and see if they are blatted.



For anyone disassembling, look in ROM 53610013 at the function @ 0x92c50. You can then cross-reference to your own ROM and find your own "special" addresses.

Notes to geeks: In this particular ROM, the calls to 0x9b29c set/clear DTCs. In this case, I'm pretty sure it's P0012 being manipulated. The table of DTCs starts at 0x94388. The way of referencing a DTC for "set/clear" action is straight from HP Lovecraft, and will send the soundest man insane with just one glance. Luckily, I'm already quite mad, so no biggie.


[edit]The key items are loaded from the 0x804xxx region. As tephra told me ages ago, 0x805xxx is the RAMBACKUP region, so 0x804xxx probably is too. Not 100% sure myself, though.


I'll update you when I have some actual data to report.


Cheers,

Rich


PS. Guess I'm assembler-bombing this thread, lol.

Wahtash

Clearing the timing chain elongation DTC does not reset the learned value. The learned value reset is a separate function.

That being said clearing a timing chain elongation DTC does not make your timing chain shorter.

richardjh

I've done some logging.

There are two pairs of values. My vehicle had...

ChainVal_1A = 3626
ChainVal_1B = 3626

ChainVal_2A = 4188
ChainVal_2B = 4188


The code works out the difference between 1A & 1B. If that is < 217, all is good.

Otherwise, it works out the difference between 2A & 2B. If that is < 307, all is good.

Otherwise, it triggers P0012.


In my case, both deltas are zero.


I ran an ECUFlash "Clear DTCs", and all four variables had the same values in them. These particular items didn't get touched.


Provisos and notes...

- All this depends on my reading of the DTC-trigger code being sane and correct. This stuff is complex, and it's easy to have too narrow a focus.

- I didn't use EvoScan DTC-clear, nor any other unit. Just the one I use in my flashing procedure.


Now, perhaps there is some other chain-stretch counter/value/history thing. I dunno. But for the P0012-triggering stuff at least, we don't seem to be impacting any ECU safety feature by reflashing and clearing DTCs. The variables it compared didn't change following a DTC clear.


Cheers,

Rich

BiGnOsE

Hi Rich,
Nice work done here ! Thanks a lot.

I'd like to ask a few more question, if I may.

First, what I understand is that

Assuming that and the fact that your deltas were 0's, is it possible that on a reset, var1A is set to const1B and var2A is set to const2B.

In that way, if you did reset the code lately, it would have been resetted and then you had 0 deltas ? and you cannot spot the change ?

Just trying to dig a little deeper here

Thanks again !

richardjh

Anything's possible.

However, I haven't flashed my RA or cleared DTCs for weeks - so the "before" values will have been kicking around for a bit.

For all we know, this thing the dealers talked about is "for information only", somewhere else, and is unrelated to the actual P0012 trigger.


Note: If anyone gets P0012, give me a shout. Let me know your ROM ID - we can check this P0012 trigger function and see what your particular variables read.

Rich

deprydation

iTrader: (1)

Anyone with a stock ROM that's been running it for quite a while might be able to pull it and we could see one that might show something.

richardjh

Stock ROMs don't support "mode23", so we won't have access to the values.

Rich

deprydation

iTrader: (1)

Eh true I'm a dumb dumb.